[Bioperl-l] Extracting gi no from refseq record

Hilmar Lapp hlapp at gnf.org
Fri Apr 4 13:44:56 EST 2003


Always email the exception message and stack trace copy&pasted, as 
otherwise no-one has a clear idea what's happened. The SeqIO swissprot 
parser will only retain the first species (and correspondingly only the 
first NCBI taxon ID). Multiple ids mean multiple secondary accession 
lines? These should all be parsed (Jason you fixed that: did you do 
that on the stable branch too?); I don't know though whether they'll 
all be indexed.

	-hilmar

On Friday, April 4, 2003, at 01:35  PM, Siddhartha Basu wrote:

> Hi,
>
> That worked like a charm, thanks a lot hilamar.
> Now i am facing another new problem. I wonder whether bioperl is able 
> to deal with the swissprot entries having multiple id lines and of 
> course multiple taxon ids. Because the Bio::Index::Swissprot module is 
> stopping on those ids having mulitple lines. Any idea how to deal with 
> these ids.
> And why the entries have such multiple lines.
>
> bye
> siddhartha
>
> Hilmar Lapp wrote:
>> The problem is not in SeqIO. The offending statement is in the method 
>>  fetch() in Bio::Index::AbstractSeq, which is the method that does 
>> the  actual retrieval work. The chunk of input is handed off to  
>> SeqIO::genbank, which I'm sure does the right job, only to be  
>> overwritten after it returns the parsed sequence object to fetch():
>>     # we essentially assumme that the primary_id for the database
>>     # is the display_id
>>     $seq->primary_id($seq->display_id()) if( defined $seq && 
>> ref($seq)  &&
>>                          $seq->isa('Bio::PrimarySeqI') );
>> Can anyone explain the usefulness of this statement? Otherwise we 
>> just  delete it and the bug is fixed. (Siddharta, as an immediate fix 
>> you  might just want to do this.)
>>     -hilmar
>
>
-- 
-------------------------------------------------------------
Hilmar Lapp                            email: lapp at gnf.org
GNF, San Diego, Ca. 92121              phone: +1-858-812-1757
-------------------------------------------------------------



More information about the Bioperl-l mailing list